AI Tool Changelog

Mar 10, 2026Major ReleaseMedium Impact

ElevenLabs Launches Dubbing Studio for Video Localization

ElevenLabs released Dubbing Studio, a comprehensive tool for automatically dubbing videos into 72 languages while preserving the original speaker's voice characteristics, timing, and emotional delivery.

•Automatic video dubbing in 72 languages
•Voice preservation across languages
•Lip-sync adjustment for dubbed content

Cursor

Cursor Launches Enterprise Plan with SSO and Audit Logs

Cursor introduced its Enterprise plan at $50/seat/month, featuring SSO, detailed audit logs, custom model routing, and the ability to use private models hosted on the customer's infrastructure.

•$50/seat/month enterprise pricing
•SSO with SAML and OIDC support
•Audit logs and compliance reporting

Mar 10, 2026Model UpdateLow Impact

ClaudevHaiku 3.5

Anthropic Updates Claude Haiku 3.5 with Faster Speed and Tool Use

Anthropic released an update to Claude Haiku 3.5, its fastest and most affordable model, improving response speed by 30% and adding tool use support. It is now the most cost-effective model with function calling.

•30% faster response times
•Tool use / function calling support added
•$0.25/M input, $1.25/M output tokens

Mar 9, 2026Model UpdateMedium Impact

Mar 8, 2026Feature UpdateHigh Impact

Perplexity Pro Upgrades to Latest Frontier Models

Perplexity updated its Pro tier to use GPT-5.2 and Claude Opus 4.6 as underlying models for search and analysis. Users can now select their preferred model for different types of queries.

•GPT-5.2 and Claude Opus 4.6 available in Pro
•Model selection per query type
•Improved citation accuracy with new models

Cursorv0.50

Cursor 0.50 Introduces Background Agents and Multi-File Edits

Cursor released version 0.50 with background agents that can work on tasks autonomously while developers focus on other work. The update also brings improved multi-file editing with better cross-file context awareness.

•Background agents for autonomous task completion
•Improved multi-file edit accuracy by 40%
•New project-wide refactoring tool

Mar 8, 2026Model UpdateMedium Impact

GeminivNano 3

Gemini Nano 3 Brings Multimodal AI to Android Devices

Google released Gemini Nano 3, an on-device model that runs entirely on Android phones without internet connectivity. It supports text, image, and audio understanding with significantly improved capabilities over Nano 2.

•Fully on-device multimodal AI (text, image, audio)
•No internet required for core features
•3x faster than Nano 2 on Pixel 9

Mar 8, 2026Feature UpdateLow Impact

Mar 7, 2026API UpdateMedium Impact

Notion AI Adds Automatic Meeting Notes from Calendar Events

Notion AI now automatically generates meeting notes pages from calendar events, including AI-generated agendas based on previous meeting context and action item tracking with assignees.

•Auto-generated meeting note pages from calendar
•AI agendas based on previous meeting context
•Action item tracking with assignees

Mar 7, 2026Feature UpdateHigh Impact

Runway Opens Gen-4 API for Developer Access

Runway launched its Gen-4 API, allowing developers to integrate 60-second video generation into their applications. Pricing starts at $0.05 per second of generated video.

•Gen-4 video generation via REST API
•Pricing at $0.05/second of video
•Webhook support for async generation

Mar 6, 2026Feature UpdateHigh Impact

ChatGPT Operator Expands to 50+ Countries with New Capabilities

OpenAI expanded ChatGPT Operator (its web-browsing agent) to 50+ countries and added new capabilities including form filling, appointment booking, price comparison shopping, and multi-site research.

•Available in 50+ countries (up from US-only)
•Automated form filling and appointment booking
•Multi-site price comparison shopping

Mar 5, 2026Feature UpdateMedium Impact

Claude Artifacts Gets Full-Stack App Preview with Databases

Anthropic upgraded Claude Artifacts to support full-stack application previews with temporary databases, backend logic, and API endpoints. Users can now prototype complete applications within Claude conversations.

•Full-stack app previews with backend logic
•Temporary SQLite databases for data persistence
•API endpoint simulation

Mar 5, 2026Feature UpdateLow Impact

ChatGPT Gets Persistent Project Memory and File Management

OpenAI added persistent project memory to ChatGPT, allowing conversations within a project to share context, files, and preferences. Users can now upload and manage files across conversations within projects.

•Persistent memory across project conversations
•File management with version history
•Project-level custom instructions

Mar 5, 2026Feature UpdateLow Impact

ElevenLabs Reader App Adds PDF and EPUB Support

ElevenLabs updated its Reader app to support PDF and EPUB files, allowing users to listen to books, research papers, and documents with natural AI narration in their preferred voice.

•PDF and EPUB file support
•Automatic chapter detection and navigation
•Adjustable reading speed (0.5x-3x)

Mar 4, 2026Major ReleaseHigh Impact

Gemini Gems Marketplace Opens with 200+ Custom AI Assistants

Google opened a public marketplace for Gemini Gems (custom AI assistants), allowing users to discover and use community-created Gems for specific tasks. Top creators can monetize their Gems.

•200+ Gems available at launch
•Creator monetization program
•Gem rating and review system

MidjourneyvV7

Midjourney V7 Launches with 3D Scene Understanding

Midjourney released V7, introducing 3D scene understanding and consistent character generation across images. The model produces significantly more photorealistic outputs with better text rendering capabilities.

•3D scene understanding for consistent spatial layouts
•Character consistency across multiple generations
•Near-perfect text rendering in images

Mar 3, 2026Feature UpdateMedium Impact

Mar 3, 2026API UpdateMedium Impact

Notion AI Launches External Data Connectors

Notion AI now connects to external data sources including Google Drive, Slack, Confluence, and Salesforce, allowing AI to search and synthesize information across all connected tools.

•Connectors for Google Drive, Slack, Confluence, Salesforce
•Cross-platform knowledge search
•Automatic sync and indexing of connected sources

Sora

OpenAI Opens Sora API Beta for Developer Video Generation

OpenAI launched the Sora API in beta, enabling developers to integrate AI video generation into their applications. The API supports text-to-video, image-to-video, and video-to-video transformations.

•Text-to-video, image-to-video, video-to-video
•Pricing at $0.10/second of generated video
•Up to 2-minute video generation via API

Mar 2, 2026Feature UpdateMedium Impact

Mar 2, 2026Feature UpdateMedium Impact

GitHub Copilot Workspace Adds Automated PR Reviews

GitHub Copilot Workspace now includes automated pull request reviews with AI-generated summaries, potential issue detection, security vulnerability scanning, and suggested improvements.

•Automated PR summaries and change descriptions
•Security vulnerability detection in code changes
•Performance impact analysis

Mar 1, 2026Major ReleaseHigh Impact

ChatGPT Search Expands to Free Users with Shopping Features

OpenAI expanded ChatGPT Search to all free users and added shopping features including product comparisons, price tracking, and direct purchase links with AI-curated recommendations.

•ChatGPT Search available on free tier
•Product comparison and price tracking
•AI-curated shopping recommendations

RunwayvGen-4

Runway Launches Gen-4 with 60-Second Video Generation

Runway released Gen-4, capable of generating up to 60 seconds of coherent video from text or image prompts. The model features dramatically improved temporal consistency and physics-aware motion.

•Up to 60 seconds of coherent video generation
•Physics-aware motion and object interaction
•Improved temporal consistency (no more morphing artifacts)

Mar 1, 2026Price ChangeMedium Impact

Feb 28, 2026Feature UpdateMedium Impact

Anthropic Launches Claude Team Plan at $30/seat/month

Anthropic introduced a new Claude Team plan priced at $30 per seat per month, positioned between Pro ($20/month) and Enterprise. It includes higher usage limits, admin controls, and shared workspace features.

•$30/seat/month with annual billing option
•5x higher usage limits than Pro
•Admin dashboard with usage analytics

February 2026

Cursorv0.49

Cursor Composer Gets Project-Wide Understanding

Cursor updated Composer with project-wide codebase understanding, allowing it to make changes that account for the full project structure, dependencies, and coding patterns across all files.

•Full project codebase understanding
•Dependency-aware code generation
•Automatic import management across files

Feb 28, 2026Feature UpdateLow Impact

Feb 27, 2026Model UpdateHigh Impact

ChatGPT Edu Plan Adds Student Safety Controls and LTI Integration

OpenAI updated ChatGPT Edu with enhanced student safety controls, LTI integration for learning management systems, and plagiarism-aware writing assistance that encourages original thinking.

•Enhanced content safety controls for students
•LTI integration with Canvas, Blackboard, Moodle
•Plagiarism-aware writing assistance mode

ClaudevOpus 4.6

Anthropic Releases Claude Opus 4.6 with Extended Thinking

Anthropic launched Claude Opus 4.6, featuring an upgraded extended thinking mode that allows the model to reason through complex problems for up to 128K tokens before responding. The model leads on SWE-bench Verified at 62.4%.

•128K token extended thinking capability
•62.4% on SWE-bench Verified (new state of the art)
•96.3% on HumanEval code generation

Feb 25, 2026Major ReleaseMedium Impact

SunovV4

Suno V4 Generates Full-Length Songs with Consistent Style

Suno released V4, capable of generating full-length 3-5 minute songs with consistent style, coherent lyrics, and professional-quality audio. The update includes multi-track editing and stem separation.

•3-5 minute full-length song generation
•Style consistency throughout entire songs
•Multi-track editing with individual stem control

Feb 25, 2026API UpdateMedium Impact

Feb 22, 2026Feature UpdateHigh Impact

Perplexity Launches Sonar API for Web-Grounded AI Responses

Perplexity released its Sonar API, enabling developers to integrate web-grounded, citation-backed AI responses into their applications. The API automatically searches the web and returns answers with source links.

•Web-grounded responses with automatic citations
•Real-time web search integrated into responses
•Pricing: $1/1K searches (Sonar) and $5/1K (Sonar Pro)

Feb 22, 2026Feature UpdateLow Impact

ChatGPT Advanced Voice Mode Gets Vision and Screen Sharing

OpenAI updated ChatGPT's Advanced Voice Mode to include real-time vision capabilities and screen sharing. Users can now show ChatGPT what they see through their camera or share their screen during voice conversations.

•Real-time camera vision during voice chats
•Screen sharing for technical assistance
•Improved voice latency (under 300ms)

Feb 20, 2026Major ReleaseHigh Impact

ElevenLabs Projects Adds Multi-Speaker and Collaboration

ElevenLabs updated its Projects feature with multi-speaker support for audiobooks and podcasts, and added collaboration tools for teams working on audio content production together.

•Multi-speaker projects with voice assignment per paragraph
•Team collaboration with shared projects
•Audio script editor with inline voice previews

GitHub CopilotvAgent Mode GA

GitHub Copilot Agent Mode Reaches General Availability

GitHub Copilot's agent mode is now generally available, allowing it to autonomously plan, implement, and test multi-file changes. The feature can create branches, run tests, and open pull requests with minimal human intervention.

•Autonomous multi-file implementation from issue descriptions
•Automatic branch creation and PR opening
•Built-in test generation and execution

Feb 20, 2026Feature UpdateMedium Impact

Feb 20, 2026Feature UpdateLow Impact

Gemini Gets Deep Google Workspace Integration Across All Apps

Google expanded Gemini's integration across all Workspace apps, adding AI-powered features in Gmail, Docs, Sheets, Slides, Meet, and Chat with cross-app context awareness.

•Cross-app context awareness (email → doc → spreadsheet)
•AI meeting summaries and action items in Meet
•Smart data analysis in Sheets with visualizations

Feb 18, 2026Feature UpdateMedium Impact

Runway Introduces Turbo Mode for 10x Faster Video Previews

Runway added a Turbo mode that generates lower-resolution video previews 10x faster, allowing creators to iterate on ideas quickly before committing to full-quality renders.

•10x faster preview generation
•480p previews in under 5 seconds
•One-click upscale to full quality

Feb 15, 2026Major ReleaseHigh Impact

Perplexity Introduces Spaces for Team Research Collaboration

Perplexity launched Spaces, a collaborative research environment where teams can share searches, build shared knowledge bases, and collaboratively explore topics with AI-assisted research threads.

•Shared research spaces for teams
•Collaborative AI research threads
•Shared knowledge base with source management

Grokv4.0

xAI Releases Grok 4 with Real-Time Information Access

xAI launched Grok 4, featuring deep integration with X (Twitter) for real-time information access and improved reasoning capabilities. The model scores 1348 on Chatbot Arena Elo, making it competitive with frontier models.

•Real-time information from X platform
•1348 Chatbot Arena Elo rating
•90.5% on MMLU benchmark

Feb 15, 2026Feature UpdateLow Impact

Feb 15, 2026Model UpdateMedium Impact

ChatGPT Enterprise Adds SCIM Provisioning and Audit Logs

OpenAI enhanced ChatGPT Enterprise with SCIM-based user provisioning, detailed audit logs for compliance, and data residency options for EU customers, addressing key enterprise security requirements.

•SCIM 2.0 for automated user provisioning
•Detailed audit logs with export capabilities
•EU data residency option

Feb 15, 2026API UpdateMedium Impact

GitHub Copilot Adds Gemini 3 and Automatic Model Routing

GitHub Copilot added Gemini 3 as a model option and introduced automatic model routing that selects the best model for each type of coding task, optimizing between speed for completions and quality for complex generation.

•Gemini 3 available alongside GPT-5.2 and Claude
•Automatic model routing by task type
•Speed-optimized completions, quality-optimized generation

Feb 14, 2026Feature UpdateMedium Impact

Gemini API Adds Google Search Grounding for Factual Accuracy

Google added Google Search grounding to the Gemini API, allowing developers to enable automatic web search verification for AI responses, significantly reducing hallucinations and providing source citations.

•Automatic Google Search grounding for responses
•Source citations included in API responses
•Configurable grounding threshold

ElevenLabsvVoice Design 2.0

ElevenLabs Launches Voice Design 2.0 with Emotion Control

ElevenLabs released Voice Design 2.0, allowing users to create custom voices by describing desired characteristics in natural language. The update adds granular emotion control with 12 emotion dimensions.

•Natural language voice creation from descriptions
•12-dimension emotion control (joy, sadness, anger, etc.)
•Real-time voice transformation during calls

Feb 12, 2026API UpdateMedium Impact

Feb 12, 2026Feature UpdateMedium Impact

Claude API Launches Batch Processing with 60% Discount

Anthropic introduced batch processing for the Claude API, allowing developers to submit large batches of requests at a 60% discount. Results are delivered within 24 hours, ideal for data processing and analysis workloads.

•60% discount on batch API requests
•Submit up to 100K requests per batch
•24-hour guaranteed turnaround

Replit

Replit Agent Adds Multi-App Workspace and Team Features

Replit updated its AI Agent to support multi-app workspaces where the agent can manage and coordinate between multiple related applications, and added team collaboration features.

•Multi-app workspace management
•Cross-app API integration by the agent
•Team collaboration with shared agent sessions

Feb 10, 2026Price ChangeMedium Impact

Feb 10, 2026Feature UpdateMedium Impact

Google Reduces Gemini Advanced Price to $14.99/month

Google reduced the Gemini Advanced subscription from $19.99 to $14.99 per month, making its most capable AI model more accessible. The price cut applies to individual plans and includes 2TB of Google One storage.

•Price reduced from $19.99 to $14.99/month
•Still includes 2TB Google One storage
•Gemini 3 Ultra access included

Feb 8, 2026Feature UpdateMedium Impact

GitHub Copilot Extensions Marketplace Launches with 100+ Tools

GitHub launched a marketplace for Copilot Extensions, allowing third-party tools to integrate directly into the Copilot experience. Launch partners include Docker, Sentry, Datadog, and LaunchDarkly.

•100+ extensions at launch
•Docker, Sentry, Datadog integrations
•Custom extension SDK for developers

Midjourney

Midjourney Web Editor Adds Inpainting and Region Selection

Midjourney updated its web editor with advanced inpainting capabilities and region-based editing. Users can select specific areas of generated images and regenerate them with new prompts while maintaining consistency.

•Brush-based inpainting for selective editing
•Region-based prompt control
•Style transfer between image regions

Feb 8, 2026Major ReleaseMedium Impact

Bolt.newv2.0

Bolt.new 2.0 Adds Backend Generation and Database Setup

Bolt.new released version 2.0 with full-stack generation capabilities, automatically setting up databases, API routes, authentication, and deployment configurations alongside frontend code.

•Full-stack app generation with databases
•Automatic authentication setup (OAuth, email)
•API route generation with documentation

Feb 7, 2026Feature UpdateMedium Impact

Sora

Sora Gets Storyboard Mode and Extended 2-Minute Videos

OpenAI updated Sora with a storyboard mode that lets users plan multi-scene videos and extended the maximum generation length to 2 minutes. The update also improves text rendering and human anatomy consistency.

•Storyboard mode for multi-scene video planning
•Extended to 2-minute video generation
•Improved text rendering in video scenes

Feb 5, 2026Feature UpdateMedium Impact

Feb 5, 2026Feature UpdateMedium Impact

Notion AI Adds Automated Workflows and Database Analysis

Notion AI received a major update adding automated workflows that can trigger actions based on database changes, and a new data analysis mode that can generate insights and visualizations from Notion databases.

•Automated workflows triggered by database changes
•AI-powered data analysis and visualization
•Smart templates that adapt to project context

Feb 3, 2026Feature UpdateMedium Impact

Claude MCP Ecosystem Grows to 500+ Integrations

Anthropic announced that the Model Context Protocol (MCP) ecosystem has grown to over 500 official integrations, making Claude the most connected AI assistant with access to databases, APIs, and enterprise tools.

•500+ official MCP integrations
•New integrations: Salesforce, SAP, Oracle
•MCP marketplace for discovering integrations

Feb 1, 2026Price ChangeMedium Impact

Runway Launches Motion Brush V2 for Precise Video Control

Runway released Motion Brush V2, an advanced tool for controlling specific motion patterns in generated videos. Users can paint motion directions, speeds, and behaviors onto specific regions of an image.

•Per-region motion control with brush strokes
•Speed and direction customization
•Looping motion for seamless GIFs

Cursor

Cursor Increases Pro Plan Price to $25/month

Cursor raised its Pro plan from $20 to $25 per month, citing increased compute costs from frontier model usage. Existing subscribers keep their current pricing for 6 months. The Business plan remains at $40/seat/month.

•Pro plan increases from $20 to $25/month
•Existing subscribers grandfathered for 6 months
•Business plan unchanged at $40/seat/month

Feb 1, 2026Feature UpdateMedium Impact

Jan 30, 2026Model UpdateMedium Impact

OpenAI Deprecates ChatGPT Plugins in Favor of GPTs and Actions

OpenAI officially deprecated the ChatGPT Plugins system, directing developers to use Custom GPTs with Actions instead. Existing plugins will continue working until April 2026, giving developers time to migrate.

•Plugins deprecated, sunset April 2026
•Custom GPTs with Actions as replacement
•Migration guide and tools provided

January 2026

Jan 30, 2026Feature UpdateMedium Impact

GitHub Copilot Adds Claude Opus 4.6 as Model Option

GitHub Copilot now allows users to select Claude Opus 4.6 as their underlying model for code completion and chat. This joins GPT-5.2 and Gemini 3 as available model options in Copilot settings.

•Claude Opus 4.6 available for code completion
•Model switching in Copilot settings
•Three model options: GPT-5.2, Claude Opus 4.6, Gemini 3

Jan 28, 2026Major ReleaseHigh Impact

ElevenLabs Conversational AI Adds Knowledge Bases

ElevenLabs updated its Conversational AI platform to support custom knowledge bases, allowing businesses to create voice agents that can answer questions from uploaded documents, websites, and databases.

•Custom knowledge base integration
•Document, website, and database ingestion
•Multi-language voice agents (72 languages)

Geminiv3 Ultra

Google Launches Gemini 3 Ultra with Native Multimodal Reasoning

Google DeepMind released Gemini 3 Ultra, its most capable model featuring native multimodal reasoning across text, image, audio, and video. The model leads on vision benchmarks and introduces real-time video understanding.

•Native multimodal reasoning across text, image, audio, video
•Real-time video understanding in Google AI Studio
•95.2% on DocVQA (new state of the art)

Jan 28, 2026Feature UpdateLow Impact

Jan 25, 2026Major ReleaseMedium Impact

Perplexity Pages Gets Visual Redesign and SEO Tools

Perplexity redesigned its Pages feature for creating AI-generated articles, adding beautiful templates, SEO optimization tools, and the ability to embed interactive charts and comparison tables.

•New visual templates for published articles
•Built-in SEO optimization suggestions
•Interactive chart and table embedding

Jan 25, 2026Feature UpdateMedium Impact

ElevenLabs Enters Music Generation with AI Sound Studio

ElevenLabs expanded beyond voice to launch AI Sound Studio, a music generation tool that creates royalty-free music tracks, sound effects, and ambient audio from text descriptions.

•Text-to-music generation up to 5 minutes
•Sound effect creation from descriptions
•Ambient audio for games and videos

Jan 22, 2026Major ReleaseHigh Impact

ChatGPT Custom GPTs Get Analytics Dashboard and Revenue Sharing

OpenAI launched an analytics dashboard for Custom GPT creators and began revenue sharing based on usage. Creators can now see detailed usage metrics and earn a share of subscription revenue.

•Detailed analytics: sessions, messages, retention
•Revenue sharing program for top GPT creators
•User feedback and rating system

Jan 22, 2026Major ReleaseMedium Impact

Claude Computer Use Reaches General Availability

Anthropic released Claude Computer Use as a generally available feature, allowing Claude to control desktop applications, navigate websites, fill out forms, and complete multi-step computer tasks autonomously.

•Full desktop application control via screenshots
•Multi-application workflow automation
•Built-in safety checks before destructive actions

v0v3.0

v0 3.0 Generates Full Next.js Apps with App Router

Vercel's v0 released version 3.0, expanding from component generation to full Next.js application scaffolding with App Router, server components, and database integration via Vercel Postgres.

•Full Next.js App Router application generation
•Server component and API route creation
•Vercel Postgres database integration

Jan 20, 2026API UpdateMedium Impact

Jan 20, 2026Feature UpdateMedium Impact

Gemini API Introduces Smart Context Caching with 75% Savings

Google launched smart context caching for the Gemini API, automatically caching repeated context across API calls. Developers save up to 75% on costs for applications that reuse system prompts or reference documents.

•Automatic context caching across API calls
•Up to 75% cost reduction on cached content
•No code changes required — automatic detection

Midjourney

Midjourney Launches Personalized Model Fine-Tuning

Midjourney introduced personalized model fine-tuning, allowing users to train the model on their preferred aesthetics by rating images. The personalized model learns individual style preferences over time.

•Image rating system to learn user preferences
•Personalized style applied to all generations
•Multiple style profiles per account

Jan 18, 2026Feature UpdateMedium Impact

Jan 18, 2026Model UpdateMedium Impact

ChatGPT Canvas Gets Real-Time Collaboration and Commenting

OpenAI updated Canvas with real-time collaboration features, allowing multiple users to work on documents and code simultaneously. The update also adds inline commenting and suggestion mode.

•Real-time multi-user collaboration
•Inline commenting and suggestions
•Version history with diff view

DeepSeekvCoder V3

DeepSeek Coder V3 Achieves 93.4% on HumanEval

DeepSeek released Coder V3, an open-source code generation model that achieves 93.4% on HumanEval, rivaling proprietary models. It specializes in code completion, generation, and debugging across 80+ languages.

•93.4% on HumanEval (rivaling proprietary models)
•Support for 80+ programming languages
•Optimized for IDE integration

Jan 18, 2026Feature UpdateLow Impact

Jan 15, 2026Feature UpdateHigh Impact

Claude Projects Adds Sharing and Collaborative Editing

Anthropic added sharing and collaboration features to Claude Projects, allowing teams to share project contexts, custom instructions, and conversation histories with granular access controls.

•Share projects with team members
•Collaborative editing of project instructions
•View-only and edit access levels

Jan 15, 2026Feature UpdateMedium Impact

Perplexity Launches Deep Research for Multi-Step Analysis

Perplexity introduced Deep Research, an agentic feature that conducts multi-step web research, synthesizes information from dozens of sources, and produces comprehensive research reports with proper citations.

•Multi-step agentic research across 50+ sources
•Automatic report generation with citations
•Comparison table generation from research data

Jan 15, 2026API UpdateMedium Impact

Runway Launches Dedicated Image Generation Model

Runway expanded beyond video to launch a dedicated image generation model that produces photorealistic images with strong text rendering, competing directly with Midjourney and DALL-E.

•Photorealistic image generation
•Accurate text rendering in images
•Seamless image-to-video pipeline

Jan 14, 2026Feature UpdateMedium Impact

OpenAI Realtime API Adds Video Input and Function Calling

OpenAI updated the Realtime API to accept video input streams alongside audio, and added function calling support. Developers can now build multimodal real-time applications with tool-use capabilities.

•Video input stream support in Realtime API
•Function calling during real-time sessions
•Reduced latency to under 250ms

Cursorv0.48

Cursor Tab Autocomplete V2 Predicts Multi-Line Edits

Cursor updated its Tab autocomplete to version 2, which now predicts multi-line edits and can suggest changes across multiple locations in a file simultaneously, significantly reducing the keystrokes needed for common refactoring tasks.

•Multi-line edit predictions in single Tab press
•Cross-location edit suggestions within a file
•50% reduction in keystrokes for refactoring

Jan 12, 2026Feature UpdateLow Impact

Jan 10, 2026Model UpdateHigh Impact

Notion AI Writing Assistant Gets Brand Voice Training

Notion AI updated its writing assistant to support custom brand voice training. Teams can upload style guides and sample content to train the AI to write in their specific brand voice and tone.

•Custom brand voice training from style guides
•Tone consistency across all AI-generated content
•Multiple brand voices per workspace

ClaudevSonnet 4

Anthropic Releases Claude Sonnet 4 — Fast and Affordable

Anthropic released Claude Sonnet 4, a mid-tier model offering 90% of Opus 4.6 performance at 1/5 the cost. It targets high-volume API use cases with an optimal balance of speed, quality, and price.

•90% of Opus 4.6 quality at 20% of the cost
•$3/M input tokens, $15/M output tokens
•2x faster response time than Opus

Jan 10, 2026Major ReleaseHigh Impact

Meta Llamav4 405B

Meta Releases Llama 4 405B with Improved Multilingual Support

Meta released Llama 4 405B, the latest iteration of its open-source LLM family. The model features significantly improved multilingual capabilities, stronger reasoning, and a new 256K context window.

•88.7% on MMLU benchmark
•256K context window (up from 128K)
•Support for 25 languages natively

Jan 8, 2026Feature UpdateMedium Impact

Jan 8, 2026Major ReleaseMedium Impact

ChatGPT Tasks Expands with Multi-Step Automations

OpenAI expanded ChatGPT Tasks to support multi-step automations that chain together web browsing, file creation, email drafts, and calendar actions. Tasks can now run on schedules with conditional logic.

•Multi-step task chains with conditional logic
•Scheduled recurring tasks (daily, weekly, custom)
•Web browsing + file creation in single tasks

Windsurfv1.0

Codeium Launches Windsurf IDE as Cursor Alternative

Codeium launched Windsurf, an AI-first IDE built on VS Code, positioning it as a more affordable alternative to Cursor. The free tier includes generous AI usage limits with support for multiple models.

•AI-first IDE built on VS Code
•Generous free tier with AI completions
•Cascade: multi-file agentic editing

Jan 7, 2026Price ChangeMedium Impact